In this dissertation, the precoded space-time block codes (STBC) schemes with limited feedback in MIMO wireless communication systems are investigated, which are classified as single mode, multimode and supermode schemes based on the number of input d...
In this dissertation, the precoded space-time block codes (STBC) schemes with limited feedback in MIMO wireless communication systems are investigated, which are classified as single mode, multimode and supermode schemes based on the number of input data streams and corresponding transmission schemes.Mode refers to the number of input data stream contained in the precoded STBC codeword. For single mode precoded STBC (SM-STBC), as the word implies only one input data stream is allowed and encoded by using one OSTBC encoder, we propose a near-optimal antenna selection precoding with cost of a limited feedback for a class of full-rate, full-diversity orthogonal STBCs constructed by serially concatenating a certain number of 2 x 2 Alamouti STBC and a constellation rotated signal vector. The proposed precoding schemes, which uses an error codebook constructed from the error vector quantization, minimizes pairwise error probability. It provides close-to-optimal performance while keeping the low complexity in spatially and temporally correlated channel.The multimode precoded STBC (MM-STBC) scheme, which adaptively varies the number of parallel 2 x 2 Alamouti OSTBC based on the current channel information, is proposed to achieve betterperformance than the multimode precoded spatial multiplexing MIMO (MM-MIMO) system for some numbers of feedback because STBC can achieve full diversity without CSI at the transmitter. Compared with the single mode precoded STBC, it can further improve the performance due to a combination of linear precoding and adaptive modulation, assuming the transmitted data-rate is independent of thenumber of data streams. Furthermore, unlike the conventional unitary precoder design which is subspace packing in Grassmann manifold to maximize the chordal distance, we proposed a unitary precoder design which minimizes average error probability directly and shows performance improvement. We also propose an optimal precoder design with non-uniform power allocation which achieves better performance than the unitary design because of power allocation. However, thedesign requires a brute-force search among thousands of precoder candidates to guarantee the optimal precoder will be selected. Thus, an approximate closed-form solution of the optimal precoder isderived in order to cut down the search complexity. It is noted that the proposed percoder design with non-uniform power allocation can be applied for the MM-MIMO system to improve performance, whichshows the effectiveness and applicability of the proposed precoder design.The supermode precoded scheme, which is a hybrid operation of the MM-MIMO and MM-STBC, allows different degrees of freedom (DOF) and different transmission depending on the current channel information and is expected to improve performance compared with the multimode precoded schemes. The system requires more accurate selection criteria to guarantee the mode of the MM-MIMO or MM-STBC providing the best performance is selected. However, the selection criteria for the multimode precoded schemes based on the minimum eigenvalue of the effective channel may not be optimal for low SNRs. Therefore, a tight bound of pairwise error probability for maximum likelihood (ML) detection is derived, which leads to new feedback allocation and selection criteria for mode adaptation in the supermode precoded scheme.